In order to increase the resolution of seismic data used to detect thin‐beds, we implemented the Spectral Inversion method (Puryear and Castagna, 2008) which examines the Thin‐Bed Tuning response in the frequency domain. Our approach included minimization of the objective function through the application of Genetic Algorithms (GA) and Simulated Annealing (SA). We tested our implementation on synthetic and real seismic data from the Barco Formation, deposited in Catatumbo basin of Colombia. Our results indicate that these two algorithms outperform traditional minimization schemes.